Display device and media resource searching method

Abstract

The invention relates to a display device and a media resource search method. The apparatus includes: a display configured to display a user interface; the controller is configured to receive a media resource search request and determine text input information of a target task; generating search prompt information based on the text input information; the search prompt information comprises a standard operation process, text input information, memory information retrieved based on the text input information and basic information for the target task, and the standard operation process is obtained by retrieving a standard operation process database by the target agent based on the text input information, the memory information and the basic information; controlling the display to display the media resources on the user interface; the media resource is obtained by executing a sub-task by the target agent, and the sub-task is obtained by calling a large model by the target agent based on the search prompt information to disassemble the target task. By adopting the scheme of the invention, the accuracy of media resource search results can be improved.

Description

Technical Field

[0001] The present application relates to the technical field of display devices, and in particular to a display device and a media resource search method. Background Art

[0002] With the rapid development of technology, more and more additional functions have emerged based on the original functions of display devices, such as the search function for media resources such as film and television resources and music resources. By integrating the media resource search function into the display device, the searched media resources can be displayed according to the user's search needs.

[0003] In traditional technology, intent recognition is usually performed based on the user's standard instructions, and then media resources are searched based on the intent recognition results. When the user's needs are complex or vague, the accuracy of intent recognition is often low, and it is impossible to search for media resources that meet the user's real needs.

[0004] Therefore, using traditional technology, there is a problem of low accuracy of searched media resources. Summary of the invention

[0005] The present application provides a display device and a media resource search method to solve the problem of low accuracy of searched media resources.

[0006] In a first aspect, some embodiments provide a display device, including:

[0007] a display configured to display content from a broadcast system or network and / or a user interface, and to display a user interface;

[0008] The controller is configured as:

[0009] Receive a media resource search request from a user, and determine text input information for describing a target task indicated by the media resource search request;

[0010] Based on the text input information, generate search prompt information for the target task; the search prompt information includes a standard operating procedure, the text input information, memory information retrieved based on the text input information, and basic information for the target task, and the standard operating procedure is obtained by the target agent by searching a standard operating procedure database based on the text input information, the memory information, and the basic information;

[0011] Control the display to display in the user interface at least one media resource searched based on the search prompt information; the at least one media resource is obtained by the target agent executing a subtask corresponding to the target task, and the subtask is obtained by the target agent calling a large model based on the search prompt information to decompose the target task.

[0012] In a second aspect, some embodiments further provide a media resource search method, the method comprising:

[0013] Receive a media resource search request from a user, and determine text input information for describing a target task indicated by the media resource search request;

[0014] Based on the text input information, generate search prompt information for the target task; the search prompt information includes a standard operating procedure, the text input information, memory information retrieved based on the text input information, and basic information for the target task, and the standard operating procedure is obtained by the target agent by searching a standard operating procedure database based on the text input information, the memory information, and the basic information;

[0015] On the user interface, at least one media resource searched based on the search prompt information is displayed; the at least one media resource is obtained by the target intelligent agent executing a subtask corresponding to the target task, and the subtask is obtained by the target intelligent agent calling a large model based on the prompt information to disassemble the target task.

[0016] In a third aspect, some embodiments further provide a media resource search device, which is applied to the display device provided in the first aspect. The device includes:

[0017] A text input information determination module, used to receive a media resource search request from a user, and determine text input information used to describe a target task indicated by the media resource search request;

[0018] A search prompt information generating module, configured to generate search prompt information for the target task based on the text input information; the search prompt information includes a standard operating procedure, the text input information, memory information retrieved based on the text input information, and basic information for the target task, wherein the standard operating procedure is obtained by the target agent by searching a standard operating procedure database based on the text input information, the memory information, and the basic information;

[0019] A media resource display module is used to display at least one media resource searched based on the search prompt information on a user interface; the at least one media resource is obtained by the target intelligent agent executing a subtask corresponding to the target task, and the subtask is obtained by the target intelligent agent calling a large model based on the prompt information to disassemble the target task.

[0020] In a fourth aspect, some embodiments further provide a display device, wherein the display device comprises a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of the media resource search method provided in the second aspect are implemented.

[0021] In a fifth aspect, some embodiments further provide a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of the media resource search method provided in the second aspect are implemented.

[0022] In a sixth aspect, some embodiments further provide a computer program product, including a computer program, which implements the steps of the media resource search method provided in the second aspect when executed by a processor.

[0023] It can be seen from the above technical solution that in some embodiments, by receiving a user's media resource search request, text input information used to describe the target task indicated by the media resource search request is determined, and based on the text input information, search prompt information for the target task is generated, and at least one media resource searched based on the search prompt information is displayed on the user interface, wherein the search prompt information includes a standard operating procedure, the text input information, memory information retrieved based on the text input information, and basic information for the target task. The standard operating procedure is obtained by the target intelligent agent by retrieving the standard operating procedure database based on the text input information, the memory information and the basic information. The media resources are obtained by the target intelligent agent executing the subtask corresponding to the target task, and the subtask is obtained by the target intelligent agent calling the large model based on the prompt information to disassemble the target task. With the help of intelligent agents to search for media resources, the intelligent agents have a certain degree of autonomy and intelligence, can interact with the environment, and can use the powerful understanding ability of the big model, so that the big model can identify complex user needs and plan tasks autonomously. In addition, since it can combine input information, memory information and basic information to retrieve the standard operating procedure database, the retrieved standard operating procedures are more adaptable to the target tasks and users, so the retrieved standard operating procedures are added to the search prompt information, which can give the big model more accurate prompts during the task decomposition process, making the searched media resources more accurate and personalized, thereby improving the user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0025] Figure 1 A schematic diagram of an operation scenario between a display device and a control device provided in some embodiments of the present application;

[0026] Figure 2 A schematic diagram of the hardware configuration of a display device provided in some embodiments of the present application;

[0027] Figure 3 A schematic diagram of the hardware configuration of a control device provided in some embodiments of the present application;

[0028] Figure 4 A schematic diagram of software configuration of a display device provided in some embodiments of the present application;

[0029] Figure 5 A schematic diagram of the configuration of various components in a display device provided in some embodiments of the present application;

[0030] Figure 6 This is a schematic diagram of obtaining memory information based on conversation history in some embodiments of the present application;

[0031] Figure 7 A schematic diagram showing the process of interaction between a display device and a server in some embodiments of the present application;

[0032] Figure 8 A schematic diagram of the steps for generating search prompt information in some embodiments of the present application;

[0033] Fig. 9 A schematic diagram showing the contents of a tool library of a display device in some embodiments of the present application;

[0034] Fig.10 A schematic diagram of the steps for constructing a standard operating procedure database in some embodiments of the present application;

[0035] Fig.11 A schematic diagram of displaying a preview control in a user interface in some embodiments of the present application;

[0036] Fig.12 A schematic diagram of playing a preview clip in a floating interface in some embodiments of the present application;

[0037] Fig.13 A timing diagram of a media resource search method is provided for some embodiments of the present application;

[0038] Fig.14 A flowchart of a media resource search method provided in some embodiments of the present application;

[0039] Fig.15 A structural block diagram of a media resource search device provided in some embodiments of the present application. DETAILED DESCRIPTION

[0040] The following embodiments are described in detail, and examples thereof are shown in the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementations described in the following embodiments do not represent all implementations consistent with the present application. They are only examples of systems and methods consistent with some aspects of the present application as detailed in the claims.

[0041] It should be noted that the brief description of terms in this application is only for the convenience of understanding the embodiments described below, and is not intended to limit the embodiments of this application. Unless otherwise specified, these terms should be understood according to their common and usual meanings.

[0042] The terms "first", "second", "third", etc. in the specification and claims of this application and the above drawings are used to distinguish similar or similar objects or entities, and do not necessarily mean to limit a specific order or sequence, unless otherwise noted. It should be understood that the terms used in this way can be interchangeable under appropriate circumstances.

[0043] The terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover but not exclude inclusion, for example, a product or device comprising a list of components is not necessarily limited to all the components expressly listed but may include other components not expressly listed or inherent to such product or device.

[0044] The term "module" refers to any known or later developed hardware, software, firmware, artificial intelligence, fuzzy logic, or combination of hardware and / or software code that is capable of performing the functions associated with that element.

[0045] In the embodiment of the present application, the display device 200 generally refers to a device with image display and data processing capabilities. For example, the display device 200 includes but is not limited to a smart TV, a mobile terminal, a computer, a monitor, an advertising screen, a wearable device, a virtual reality device, an augmented reality device, etc.

[0046] Figure 1 This is a schematic diagram of an operation scenario between a display device and a control device provided in some embodiments of the present application. Figure 1As shown in FIG. 1 , the user can operate the display device 200 through touch operation, the mobile terminal 300 and the control device 100. For example, the control device 100 can be a remote controller, a stylus pen, a handle, etc.

[0047] The mobile terminal 300 can be used as a control device for performing human-computer interaction between the user and the display device 200. The mobile terminal 300 can also be used as a communication device for establishing a communication connection with the display device 200 and performing data interaction. In some embodiments, the mobile terminal 300 can install software applications with the display device 200, and achieve connection and communication through a network communication protocol to achieve the purpose of one-to-one control operation and data communication. The audio and video content displayed on the mobile terminal 300 can also be transmitted to the display device 200 to achieve a synchronous display function.

[0048] like Figure 1 As also shown in FIG. 4 , the display device 200 also communicates data with the server 400 through various communication methods. The display device 200 may be allowed to communicate and connect through a local area network (LAN), a wireless local area network (WLAN) and other networks.

[0049] The display device 200 may provide a broadcast receiving television function, and may also additionally provide an intelligent network television function of a computer support function, including but not limited to network television, smart television, Internet Protocol television (IPTV), etc.

[0050] Figure 2 Some embodiments of the present application provide Figure 1 2 is a block diagram of the hardware configuration of the display device 200.

[0051] In some embodiments, the display device 200 may include at least one of a tuner 210, a communication device 220, a detector 230, a device interface 240, a controller 250, a display 260, an audio output device 270, a memory, a power supply, and a user input interface.

[0052] In some embodiments, the detector 230 is used to collect signals of the external environment or external interaction. For example, the detector 230 includes a light receiver, a sensor for collecting the intensity of ambient light; or, the detector 230 includes an image collector, such as a camera, which can be used to collect external environment scenes, user attributes or user interaction gestures; or, the detector 230 includes a sound collector, such as a microphone, etc., for receiving external sounds.

[0053] In some embodiments, the display 260 includes a display function component for presenting a picture, and a driving component for driving an image display. The display 260 is used to receive an image signal output from the controller 250 for display. For example, the display 260 can be used to display video content, image content, and components of a menu control interface and a user control UI interface.

[0054] In some embodiments, the communication device 220 is a component for communicating with an external device or server 400 according to various communication protocol types. The display device 200 may be provided with a plurality of communication devices 220 according to different supported communication modes. For example, when the display device 200 supports wireless network communication, the display device 200 may be provided with a communication device 220 including a WiFi function. When the display device 200 supports Bluetooth connection communication, the display device 200 needs to be provided with a communication device 220 including a Bluetooth function.

[0055] The communication device 220 can enable the display device 200 to communicate with the external device or server 400 by wireless or wired connection. Among them, the wired connection can connect the display device 200 with the external device through components such as data cables and interfaces. The wireless connection can connect the display device 200 with the external device through wireless signals or wireless networks. The display device 200 can establish a connection relationship with the external device directly, or indirectly establish a connection relationship through a gateway, a router, a connection device, etc.

[0056] In some embodiments, the controller 250 may include at least one of a central processing unit, a video processor, an audio processor, a graphics processor, and a power processor, and a first interface to an nth interface for input / output. The controller 250 controls the operation of the display device and responds to the user's operation through various software control programs stored in the memory. The controller 250 controls the overall operation of the display device 200.

[0057] In some embodiments, the controller 250 and the tuner-demodulator 210 may be located in different separate devices, that is, the tuner-demodulator 210 may also be located in an external device of the main device where the controller 250 is located, such as an external set-top box.

[0058] In some embodiments, the user may input a user command through a graphical user interface (GUI) displayed on the display 260 , and the user input interface receives the user input command through the graphical user interface (GUI).

[0059] In some embodiments, the audio output device 270 may be a local speaker of the display device 200, or may be an external audio output device of the display device 200. In particular, for the external audio output device of the display device 200, the display device 200 may also be provided with an external audio output terminal, and the audio output device may be connected to the display device 200 through the external audio output terminal to output the sound of the display device 200.

[0060] In some embodiments, the user input interface 280 may be used to receive instructions from a user.

[0061] Figure 3 Some embodiments of the present application provide Figure 1 The hardware configuration diagram of the control device in the figure. Figure 3 As shown, the control device 100 may include: a controller 110, a communication interface 130, a user input / output interface, a memory, and a power supply.

[0062] The control device 100 is configured to control the display device 200 , and can receive user input operation instructions, and convert the operation instructions into instructions that the display device 200 can recognize and respond to, playing the role of an interactive intermediary between the user and the display device 200 .

[0063] In some embodiments, the control device 100 may be a smart device, for example, the control device 100 may be installed with various applications for controlling the display device 200 according to user needs.

[0064] In some embodiments, Figure 1 As shown, the mobile terminal 300 or other intelligent electronic devices can play a similar function as the control device 100 after installing the application for controlling the display device 200 .

[0065] The controller 110 includes a processor 112, a RAM 113, a ROM 114, a communication interface 130, and a communication bus. The controller 110 is used to control the operation and operation of the control device 100, as well as the communication and cooperation between the internal components and the external and internal data processing functions.

[0066] The communication interface 130 implements communication of control signals and data signals with the display device 200 under the control of the controller 110. The communication interface 130 may include at least one of other near field communication modules such as a WiFi chip 131, a Bluetooth module 132, and an NFC module 133.

[0067] The user input / output interface 140 , wherein the input interface includes at least one of other input interfaces such as a microphone 141 , a touch panel 142 , a sensor 143 , and a button 144 .

[0068] In some embodiments, the control device 100 includes at least one of a communication interface 130 and an input / output interface 140. The control device 100 is configured with a communication interface 130, such as a WiFi, Bluetooth, NFC or other module, and can encode the user input command through the WiFi protocol, Bluetooth protocol, or NFC protocol and send it to the display device 200.

[0069] The memory 190 is used to store various operating programs, data and applications for driving and controlling the control device 100 under the control of the controller. The memory 190 can store various control signal instructions input by the user.

[0070] The power supply 180 is used to provide operating power support for each component of the control device 100 under the control of the controller.

[0071] In order to perform user interaction, in some embodiments, the display device 200 may run an operating system. The operating system is a computer program for managing and controlling hardware resources and software resources in the display device 200. The operating system may provide a user interface (control the display device), allow the user to interact with the display device 200, and support the running of various application programs.

[0072] It should be noted that the operating system may be a native operating system based on a specific operating platform, or a third-party operating system deeply customized based on a specific operating platform, or an independent operating system specially developed for the display device.

[0073] The operating system can be divided into different modules or layers according to the functions implemented, such as Figure 4 As shown, in some embodiments, the system is divided into four layers, from top to bottom, namely, the application layer (Applications layer) (referred to as "application layer"), the application framework layer (Application Framework layer) (referred to as "framework layer"), the system library layer and the kernel layer.

[0074] In some embodiments, the application layer is used to provide services and interfaces for applications so that the display device 200 can run applications and interact with users based on applications. At least one application can be run in the application layer, and these applications can be window programs, system settings programs, clock programs, etc. that come with the operating system; they can also be applications developed by third-party developers. In specific implementations, the application packages in the application layer are not limited to the above examples.

[0075] The framework layer provides application programming interfaces (APIs) and programming frameworks for applications. The application framework layer includes some predefined functions. The application framework layer is equivalent to a processing center that determines the actions of applications in the application layer. Applications can access system resources and obtain system services during execution through the API interface.

[0076] like Figure 4 As shown, in the embodiment of the present application, the application framework layer includes a view system, managers, content providers, etc., wherein the view system can design and implement the interface and interaction of the application, and the view system includes lists, grids, text boxes, buttons, etc. The manager includes at least one of the following modules: an activity manager is used to interact with all activities running in the system; a location manager is used to provide system services or applications with access to system location services; a package manager is used to retrieve various information related to the application package currently installed on the device; a notification manager is used to control the display and clearing of notification messages; and a window manager is used to manage icons, windows, toolbars, wallpapers, and desktop components on the user interface.

[0077] In some embodiments, the activity manager is used to manage the life cycle of each application and the usual navigation back function, such as controlling the exit, opening, and back of the application. The window manager is used to manage all window programs, such as obtaining the display screen size, determining whether there is a status bar, locking the screen, capturing the screen, and controlling the display window changes, for example, reducing the display window, shaking the display, distorting the display, etc.

[0078] In some embodiments, the system runtime layer can provide support for the framework layer. When the framework layer is used, the operating system will run the instruction library contained in the system runtime layer, such as the C / C++ instruction library, to implement the functions to be implemented by the framework layer.

[0079] In some embodiments, the kernel layer is a functional layer between the hardware and software of the display device 200. The kernel layer can implement functions such as hardware abstraction, multitasking, and memory management. Figure 4As shown, the kernel layer may be configured with hardware drivers, and the drivers included in the kernel layer may be at least one of the following drivers: audio driver, display driver, Bluetooth driver, camera driver, WIFI driver, USB driver, HDMI driver, sensor driver (such as fingerprint sensor, temperature sensor, pressure sensor, etc.), and power driver, etc.

[0080] It should be noted that the above example is only a simple division of the operating system functions and does not constitute a limitation on the specific operating system form of the display device 200 in the embodiment of the present application. Depending on factors such as the function of the display device and the type of operating system, the number of levels and specific level types contained in the operating system may be expressed in other forms.

[0081] In some embodiments, a display device is provided, referring to Figure 5 The display device includes: a display 502, configured to display content from a broadcast system or network and / or a user interface, and to display a user interface; a controller 504, configured to: receive a user's media resource search request, determine text input information used to describe a target task indicated by the media resource search request; generate search prompt information for the target task based on the text input information; and control the display to display at least one media resource searched based on the search prompt information in the user interface.

[0082] Among them, the user interface (UI) refers to the medium for interaction and information exchange between the system and the user, which can include various forms such as graphical user interface (GUI), command line interface (CLI), touch interface, and voice user interface (VUI).

[0083] The media resource search request is used to request a search for media resources, and the media resources may be any one of film and television resources, audio resources, text resources, game resources, etc. Exemplarily, the user may input text through the control device to trigger the media resource search request. For example, the user may input text input information of "play a movie for me" through the control device, and trigger the media resource search request through a confirmation operation. Exemplarily, the display device may also include a sound collector configured to obtain user voice, so that the user may trigger the media resource search request through voice input. The target task indicated by the media resource search request is the media resource search task indicated by the media resource search request. The text input information is text information describing the media resource search task. When the user triggers the media resource search request through text input, the text input information is the text information input by the user through text input. When the user triggers the media resource search request through voice input, the text input information is the voice information obtained by converting the voice input by the user through voice input into text.

[0084] The search prompt information includes standard operating procedures, text input information, memory information retrieved based on the text input information, and basic information for the target task. The standard operating procedures are obtained by the target agent by searching the standard operating procedure database based on the text input information, memory information, and basic information. Among them:

[0085] The basic information for the target task is current information related to the target task, such as the current time and the user's location, user gender, age, etc.

[0086] Standard Operating Procedure (SOP) is a description of the standard operating steps and requirements of an event in a unified format, which can be used to guide the task execution of the agent. There are multiple standard operating procedures stored in the retrieval standard operating procedure database. Optionally, the standard operating procedures in the database can be formulated by experts based on experience.

[0087] The target agent refers to the AI ​​(Artificial Intelligence) agent used for media resource search. The AI ​​agent is a system driven by a large language model as its brain. It has the ability to autonomously understand, perceive, plan, remember and use tools, and can automatically perform complex tasks. Unlike traditional artificial intelligence, the AI ​​agent has the ability to gradually complete a given goal by thinking independently and calling tools. This agent can perceive the environment, make decisions and perform actions. It has autonomy and adaptability. It can rely on the capabilities given by AI to complete specific tasks and continuously improve itself in the process. It can be imagined as a super smart robot assistant with a powerful brain and learning ability. It can not only understand human language, but also improve its skills in a certain field by learning and analyzing data.

[0088] In some exemplary embodiments, the target intelligent agent retrieves the standard operating procedure database based on text input information, memory information and basic information. Specifically, the target intelligent agent generates a target vector based on text input information, memory information and basic information, performs a vector similarity search in the standard operating procedure database, and obtains one or more standard operating procedures with the highest similarity to the target vector.

[0089] The memory information is retrieved from the memory storage based on the text input information, and is the memory related to the target task. The memory information is the historical information related to the target task. In some embodiments, the memory information in the memory storage can be generated by the following steps:

[0090] 1. Input: User’s historical conversation context.

[0091] 2. Conversation summary: Conversation context, based on large model and few sample prompts, generates summary information. Including: conversation summary, topic, personality tag, interest tag, etc.

[0092] 3. Memory storage: Conversation summaries, key events, and topics are stored in the vector library, and personality tags and interest tags are stored in MySQL.

[0093] 4. Memory update: including adding, merging, and forgetting. The input is new memory and historical memory, and different prompts are used to let the large model merge and forget memories.

[0094] For example, see Figure 6 , which is a schematic diagram of obtaining memory information based on conversation history in some exemplary embodiments. Figure 6 It can be seen that through the conversation history, we can obtain conversation summaries such as topics, conversation summaries, as well as tags and basic information.

[0095] At least one media resource is obtained by the target intelligent agent executing a subtask corresponding to the target task. The subtask is obtained by the target intelligent agent calling a large model based on search prompt information to decompose the target task.

[0096] In some optional embodiments, the target agent executes the subtask corresponding to the target task by calling the tool interface indicated by the subtask. In this case, the search prompt information input by the target agent to the large model may also include tool description information of all available tools, so that the large model can return the decomposed subtasks and the tool interface indicated by them when decomposing the task.

[0097] In some optional embodiments, the large model called by the target agent is a large language model (LLM) that has been trained through professional knowledge learning in the field of media resource search and instruction fine-tuning.

[0098] Specifically, a user can trigger a media resource search request through voice input or text input, so that the controller of the display device can receive the user's media resource search request, and then determine the text input information of the target task indicated by the media resource search request. The controller can provide the text input information to the target intelligent agent, and the target intelligent agent can obtain memory information based on the text input information, and further obtain basic information for the target task. Based on the text input information, memory information and basic information, the standard operating procedure database is retrieved to obtain the standard operating procedure, and then search prompt information can be constructed based on the standard operating procedure, text input information, memory information retrieved based on the text input information, and basic information for the target task.

[0099] The target agent can input the search prompt information to the large model, and the large model decomposes the target task into subtasks, and then the target agent performs the subtask corresponding to the target task to obtain at least one media resource. After obtaining at least one media resource, the controller can control the display device to display these media resources on the user interface displayed by it.

[0100] In an optional embodiment, the target agent and the large model can be deployed on the server, which provides service support for multiple display devices, and performs media resource search through interaction between the display devices and the server. Figure 7 Specifically, after receiving the user's media resource search request and determining the text input information used to describe the target task indicated by the media resource search request, the display device can send the text input information to the server, and the server generates search prompt information for the target task through the target agent based on the text input information. The target agent can obtain memory information from the memory storage, obtain tool description information from the tool library, and call the large model to decompose the target task to obtain subtasks, and then obtain at least one media resource by executing the subtask corresponding to the target task. The server can return the media resource to the display device, and then the display device can control the display to display the media resource in the user interface.

[0101] In an optional embodiment, when the computing power of the display device meets the requirements, the target agent and the large model can also be deployed on the display device, and the media resource search and display can be completed by the display device alone.

[0102] The above technical scheme has the following advantages or beneficial effects: receiving a user's media resource search request, determining text input information used to describe the target task indicated by the media resource search request, generating search prompt information for the target task based on the text input information, and displaying at least one media resource searched based on the search prompt information on a user interface, wherein the search prompt information includes a standard operating procedure, text input information, memory information retrieved based on the text input information, and basic information for the target task; the standard operating procedure is obtained by the target intelligent agent by retrieving a standard operating procedure database based on the text input information, memory information, and basic information; the media resources are obtained by the target intelligent agent executing a subtask corresponding to the target task; the subtask is obtained by the target intelligent agent calling a large model based on the prompt information to decompose the target task. With the help of intelligent agents to search for media resources, the intelligent agents have a certain degree of autonomy and intelligence, can interact with the environment, and can use the powerful understanding ability of the big model, so that the big model can identify complex user needs and plan tasks autonomously. In addition, since it can combine input information, memory information and basic information to retrieve the standard operating procedure database, the retrieved standard operating procedures are more adaptable to the target tasks and users, so the retrieved standard operating procedures are added to the search prompt information, which can give the big model more accurate prompts during the task decomposition process, making the searched media resources more accurate and personalized, thereby improving the user experience.

[0103] In some embodiments, reference Figure 8 When the controller executes the step of generating search prompt information for the target task based on the text input information, the controller is further configured to execute steps 802 to 806, wherein:

[0104] Step 802: Obtain memory information for the target task based on the text input information through the target agent, and obtain basic information for the target task.

[0105] Specifically, the text input information describes the target task, and the target agent can generate a corresponding vector based on the text input information. The memory information in the memory storage is generated according to historical conversations. The historical conversations contain text input information of historical tasks. The text input information of historical tasks can be stored corresponding to the memory information generated according to the historical conversations of the historical tasks. Therefore, the target agent can retrieve the text input information of at least one most similar historical task from the memory storage through vector similarity, and use the memory information corresponding to the text input information of at least one most similar historical task as the memory information for the target task.

[0106] Furthermore, the target agent can also obtain basic information for the target task, for example, it can obtain the current time by querying the clock module of the display device and obtain the current position by querying the positioning module of the display device.

[0107] Step 804, through the target intelligent agent, based on the text input information, memory information and basic information, the standard operating procedure database is retrieved to obtain the standard operating procedure for the target task.

[0108] Specifically, the standard operating procedures in the standard operating procedure database can be generated based on the execution trajectory of the agent in the historical tasks. The standard operating procedures can be stored corresponding to the text input information, memory information and basic information of the historical tasks used to generate the standard operating procedures. Therefore, when retrieving the standard operating procedures for the target task, the target agent can splice the text input information, memory information and basic information to obtain comprehensive information, generate a vector for the comprehensive information, and then use the generated vector to retrieve the standard operating procedure database to obtain the standard operating procedure for the target task.

[0109] Step 806, through the target agent, the text input information, memory information, basic information and standard operating procedures are spliced ​​to obtain search prompt information for the target task.

[0110] Specifically, the target agent can splice text input information, memory information, basic information and standard operating procedures to obtain search prompt information for the target task.

[0111] In the above embodiment, when the intelligent agent retrieves the standard operating procedures for the target task, the retrieved standard operating procedures can be more compatible with the target task and the user based on text input information, memory information and basic information, thereby generating more accurate search prompt information.

[0112] In some embodiments, when the controller executes the step of obtaining basic information for the target task, it is further configured to: determine, through the target intelligent agent, multiple candidate basic information that can be currently obtained and their respective definitions, and obtain historical text input information and historical basic information related to the target task from the memory information; through the target intelligent agent, input the prompt information generated by splicing the text input information, multiple candidate basic information and their respective definitions, historical text input information and historical basic information into a large model to obtain the basic information for the target task.

[0113] Among them, the multiple candidate basic information currently available refers to all the basic information currently available to the target agent, and the definition of the candidate basic information is the user's interpretation of the candidate basic information. The historical text input information related to the target task refers to the text input information of similar tasks to the target task stored in the memory storage, and the historical basic information related to the target task refers to the basic information of similar tasks to the target task stored in the memory storage. Each similar task corresponding to the target task can correspond to a historical text input information and one or more basic information.

[0114] Specifically, considering that there are many dimensions of basic information that can be obtained, if the information is obtained indiscriminately, the time cost and computing cost will be high. In some embodiments, after the target intelligent agent has determined the multiple candidate basic information and their respective definitions that are currently available, it can retrieve the memory storage based on the text input information corresponding to the target task, and determine the similar tasks corresponding to the target task, and then obtain the text input information and basic information corresponding to the similar tasks, and obtain the historical text input information and historical basic information related to the target task. The target intelligent agent can further generate prompt information based on the text input information, multiple candidate basic information and their respective definitions, historical text input information and historical basic information by splicing, and input the prompt information into the large model, and then the large model can output the basic information required for the target task.

[0115] In the above embodiment, the prompt information generated by splicing text input information, multiple candidate basic information and their respective definitions, historical text input information and historical basic information is input into the big model. The big model can make decisions based on the currently available information and the basic information of historical tasks, and output the basic information required for the target task, which can save time and computing costs.

[0116] In some embodiments, the target agent executes the subtask by calling the tool interface indicated by the subtask, and the controller is further configured to: retrieve at least one tool description information from a tool description information database based on text input information through the target agent; when the controller executes the step of splicing text input information, memory information, basic information and standard operating procedures through the target agent to obtain search prompt information for the target task, the controller is further configured to: splice text input information, memory information, basic information, standard operating procedures and at least one tool description information through the target agent to obtain search prompt information for the target task.

[0117] Among them, the tool description information is the basic information describing the tool, such as the Chinese and English names of the tool, the tool interface address, input parameters and output parameters, etc., for the intelligent agent to understand and call.

[0118] Specifically, when the target intelligent agent executes the subtask obtained by disassembling the large model, it can execute it by calling the tool interface indicated by the subtask. Therefore, when the large model disassembles the target task to obtain the subtask, it can also return tool information for the subtask that needs to call the tool interface. That is, the large model outputs to the intelligent agent the instruction information on how to execute the next action, so that the intelligent agent can accurately call the tool interface to obtain information or execute control.

[0119] In order to include tool information in the output information of the large model, the target agent can also add tool description information when generating search prompt information, that is, the target agent can splice text input information, memory information, basic information, standard operating procedures and tool description information to obtain search prompt information for the target task. Exemplarily, the tool description information here can be the tool description information of all available tools. In some other embodiments, in order to further reduce the amount of calculation, the target agent can also filter the tool description information first, so that the search prompt information for the target task can be obtained based on at least one tool description information, text input information, memory information, basic information and standard operating procedures obtained by filtering.

[0120] Exemplarily, in the tool description information database, when screening tool description information, the target agent can search from the tool description information database based on text input information to obtain at least one tool description information for the target task.

[0121] In some exemplary embodiments, reference Fig. 9 The tool library of the display device may include tools for various scenarios such as food, movie watching, and fitness. Taking the movie watching scenario as an example, in this scenario, the media resource search request is a film and television resource search request. The available tool sets may include, for example, "chasing dramas", "recommendations", and "atmosphere creation". Taking the "atmosphere creation" tool set as an example, this tool set may include tools such as volume control and lighting adjustment.

[0122] In some exemplary embodiments, in the movie-watching scenario, the process of generating tool description information is as follows: 1. Native tool API interface development: implement atomic functions, such as film and television search: search for movies based on title, actor, director and other information; volume setting: increase or decrease the sound of the device; weather query: query the corresponding weather information based on time and location, etc. 2. Tool platform proxy service access: unify the management of native interfaces for easy upper-level calls. 3. Tool description schema writing: describe the basic information of the tool, such as the Chinese and English names of the tool, the tool interface address, input and output parameters, etc., for the intelligent agent to understand and call. The target intelligent agent is based on the schema file, that is, the intelligent agent calling tool class can be dynamically created.

[0123] In the above embodiment, since the tool description information can be added to the search prompt information, the large model can be prompted to output accurate tool calling information, and the tool description information is retrieved based on the text input information, which can reduce the calculation amount of the large model.

[0124] In some embodiments, reference Fig.10 The controller is further configured to execute the step of building a standard operating procedure database, which includes steps 1002 to 1008, wherein:

[0125] Step 1002, for each historical task in the historical task set, obtain the text input information, agent execution trajectory and searched media resources corresponding to the historical task.

[0126] The historical task set includes multiple historical tasks. A historical task is a media resource search task before a target task. A historical task can be triggered by the user corresponding to the target task through text input information, or by other users. The text input information corresponding to the historical task is used to describe the historical task. The agent execution trajectory refers to the trajectory composed of the execution actions of the agent in the process of executing the decomposed subtasks.

[0127] Specifically, for each historical task in the historical task set, the controller can obtain the text input information, the agent execution trajectory, and the searched media resources of the historical task from the database.

[0128] Step 1004, construct a reflective agent, and determine the execution result of the historical task based on the text input information of the historical task, the agent execution trajectory and the searched media resources through the reflective agent.

[0129] Among them, the execution result of a historical task can be a result used to characterize the successful execution of the historical task or a result used to characterize the failed execution of the historical task. If the execution result of a historical task is a result used to characterize the successful execution of the historical task, then the historical task is a successful task. If the execution result of a historical task is a result used to characterize the failed execution of the historical task, then the historical task is a failed task.

[0130] Specifically, the controller can construct a reflective agent. For historical tasks in the historical task set, the reflective agent can splice the text input information of the historical task, the agent execution trajectory and the searched media resources to obtain prompt information. Through the prompt information, the large model is called to judge whether the historical task is executed successfully or not, and the execution result of the historical task is obtained.

[0131] Step 1006 , filtering successful tasks whose execution results indicate successful execution from the historical task set, and calling the large model to generate a first standard operating procedure based on the filtered successful tasks.

[0132] Step 1008: construct a standard operating procedure database based on the generated first standard operating procedure.

[0133] Specifically, the controller can filter out successful tasks from the historical task set, and then call the large model to generate the first standard operating procedure based on the filtered successful tasks. Exemplarily, the controller can generate a corresponding first standard operating procedure for each successful task. Specifically, for each historical task, the controller can obtain the text input information, agent thinking information, and calling tool information corresponding to the historical task to obtain prompt information to prompt the large model to generate the first standard operating procedure corresponding to the successful task.

[0134] The controller can further store the generated first standard operating procedure and its corresponding text input information in a database to construct a standard operating procedure database, wherein the text input information corresponding to the first standard operating procedure is used to generate text input information corresponding to the historical tasks of the first standard operating procedure.

[0135] In the above embodiment, since a reflective agent can be constructed, the reflective agent can screen successful tasks based on the text input information of historical tasks, the agent execution trajectory and the searched media resources, and then call the large model based on the successful tasks to generate standard operating procedures. In this way, high-quality standard operating procedures can be generated, and the efficiency of generating standard operating procedures can be improved, thereby improving the construction efficiency and quality of the standard operating procedure database.

[0136] In some embodiments, when the controller executes the step of calling the big model to generate the first standard operating procedure based on the filtered successful tasks, it is further configured to: for each filtered successful task, based on the media resource search request corresponding to each successful task, group and cluster each successful task to obtain multiple cluster clusters; for each cluster cluster, from the successful tasks included in the cluster cluster, filter the successful tasks whose number of subtasks is less than a preset threshold to obtain high-quality successful tasks; for each high-quality successful task, splice the text input information, agent thinking information and calling tool information corresponding to the high-quality successful task to obtain the first prompt information, so as to prompt the big model to generate the first standard operating procedure corresponding to the high-quality successful task.

[0137] The agent thinking information refers to the agent's self-thinking information during the execution of the subtask, and the tool calling information is information used to characterize the tools called by the agent during the execution of the subtask.

[0138] Specifically, for the multiple successful tasks screened out, the controller can group and cluster these successful tasks based on the media resource search requests corresponding to these successful tasks, and group the successful tasks corresponding to the same or similar media resource search requests into one category, thereby obtaining multiple clusters. Exemplarily, the controller can determine the similarity between the media resource search requests by calculating the text similarity between the text input information corresponding to the media resource search requests, and then group the successful tasks corresponding to the media resource search requests with similarity greater than a preset threshold into one category.

[0139] Furthermore, for each cluster, the controller can filter out high-quality successful tasks from the successful tasks contained in the cluster. A high-quality successful task is a successful task whose number of subtasks obtained by decomposition is less than a preset threshold. In each cluster, other successful tasks except high-quality successful tasks are non-high-quality successful tasks.

[0140] After determining the high-quality successful tasks, for each high-quality successful task, the controller can splice the text input information, agent thinking information and calling tool information corresponding to the high-quality successful task to obtain the first prompt information, and input the first prompt information into the large model, so as to prompt the large model to generate the first standard operating procedure corresponding to the high-quality successful task.

[0141] In the above embodiment, since high-quality tasks can be further screened out from successful tasks and a labeling operation process is generated for the high-quality tasks, the data quality in the labeling operation process database is further improved.

[0142] In some embodiments, the controller is further configured to: for each non-high-quality successful task in each cluster, splice the text input information, agent thinking information, tool calling information and tool return results corresponding to the non-high-quality successful task to obtain a second prompt information; input the second prompt information into the large model to prompt the large model to optimize the sub-tasks of the non-high-quality successful task, and then generate a second standard operating procedure based on the optimized sub-tasks; when the controller constructs a standard operating procedure database based on the generated first standard operating procedure, it is further configured to: construct a standard operating procedure database based on the generated first standard operating procedure and the generated second standard operating procedure.

[0143] Specifically, in this embodiment, a standard operating procedure can also be generated based on non-high-quality successful tasks to enrich the amount of data in the standard operating procedure database. Specifically, for each non-high-quality successful task, the controller can obtain the text input information, agent thinking information, tool calling information and tool return results corresponding to the non-high-quality successful task, and then splice the obtained information to generate prompt information. The controller further inputs the generated prompt information into the large model to prompt the large model to optimize the subtasks of the non-high-quality successful task, and generate a second standard operating procedure based on the optimized subtasks. Exemplarily, subtask optimization can be, for example, removing invalid / erroneous subtasks, modifying subtasks, etc.

[0144] Furthermore, when constructing the standard operating procedure database, the controller can construct the standard operating procedure database based on the generated first standard operating procedure and the generated second standard operating procedure. In the standard operating procedure database, each standard operating procedure and its corresponding text input information are stored accordingly.

[0145] In this embodiment, since the subtasks of non-high-quality successful tasks can be optimized through the large model, a second standard operating procedure is generated based on the optimized subtasks. By generating standard operating procedures for non-high-quality successful tasks, the data in the standard operating procedure database can be enriched.

[0146] In some embodiments, the search prompt information also includes negative prompt information retrieved from the standard operating procedure database, and the controller is further configured to: for a failed task, concatenate the text input information, agent thinking information, tool calling information and tool return results corresponding to the failed task to obtain a third prompt information to prompt the large model to determine the first subtask that has no positive effect on the failed task; for a non-high-quality successful task, concatenate the text input information, agent thinking information, tool calling information and tool return results corresponding to the failed task to obtain a fourth prompt information to prompt the large model to determine the second subtask that has no positive effect on the non-high-quality successful task; generate negative prompt information based on the first subtask and the second subtask, and store it in the standard operating procedure database.

[0147] In this embodiment, for failed tasks, the text input information, agent thinking information, tool calling information and tool return result corresponding to the failed task can be spliced ​​together to generate prompt information, and the prompt information can be input into the large model. The large model can determine the first subtask that has no positive effect on the failed task, and then generate negative prompt information based on the first subtask.

[0148] For non-high-quality successful tasks, the text input information, agent thinking information, tool calling information and tool return results corresponding to the failed tasks can be spliced ​​together to generate prompt information, and the prompt information can be input into the large model. The large model can determine the second subtask that has no positive effect on the non-high-quality successful task, and then generate negative prompt information based on the second subtask.

[0149] The generated negative prompt information can be stored in the standard operating procedure database, and stored corresponding to the text input information corresponding to the task used to generate the negative prompt information. Then, when the target intelligent agent searches the standard operating procedure database, the negative prompt information can also be retrieved. Adding the negative prompt information to the search prompt information of the large model can further improve the understanding ability of the large model, thereby obtaining more accurate media resources.

[0150] In some embodiments, the media resource search request is a request to search for film and television resources, at least one media resource is a film and television resource, and the controller is further configured to: control the display to display a preview control corresponding to each film and television resource in the user interface; in response to a user trigger operation on the preview control, control the display to display a floating page in the user interface; control the display to play a preview clip of the film and television resource in the floating page.

[0151] The preview segment may be one of the segments in the film and television resource, for example, a popular segment in the film and television resource, or the preview segment may be a segment obtained by splicing multiple segments in the film and television resource.

[0152] Specifically, in this embodiment, when the controller controls the display to display film and television resources in the user interface, it can also control the display to display the preview control corresponding to each film and television resource. The user can trigger the preview control, for example, click the preview control, so that the controller can respond to the trigger operation and control the display to display a floating page in the user interface. The controller can further obtain a preview clip of the film and television resource, and then control the display to play the preview clip in the floating page.

[0153] For example, see Fig.11 and Fig.12 , each video resource displayed on the display interface has a corresponding preview control, such as Fig.11 The video resource 1102 is displayed with a preview control 1104. When the user clicks the preview control, a floating interface can be displayed and a preview clip can be played on the floating interface. For example, when the user clicks the preview control 1104, Fig.12 Play the preview clip in the floating interface.

[0154] In the above embodiment, since the user interface can be controlled to display a floating page and a preview clip on the floating page in response to the user's triggering operation on the preview control, the user can determine whether he is interested in the film and television resource by watching the preview clip, thereby further improving the user experience.

[0155] In some embodiments, the controller is further configured to: obtain user feedback information regarding at least one media resource; if the feedback information indicates that the user is dissatisfied with at least one media resource, obtain supplementary information input by the user; and update the media resources displayed on the user interface according to the supplementary information.

[0156] The supplementary information may be information that further supplements the description of the media resource search request, such as further detailed description or content that is not desired to be searched.

[0157] After the searched media resources are displayed, the searched media resources can be continuously adjusted through the interaction between the user and the display device until the media resources that satisfy the user are obtained.

[0158] Specifically, the controller may obtain feedback information from the user for at least one displayed media resource. The feedback information may be text or voice, or may be information formed by triggering a selection operation on a feedback control. The feedback control may include a "satisfied" control and a "dissatisfied" control. The feedback information may indicate that the user is satisfied with the searched media resource, or may indicate that the user is dissatisfied with the searched media resource.

[0159] When the feedback information indicates that the user is satisfied with the searched media resources, the controller may end the current media resource search and store the searched media resources in the media resource database for subsequent reference.

[0160] When the feedback information indicates that the user is not satisfied with the searched media resources, the controller can obtain the supplementary information input by the user through voice interaction, text interaction, or touch screen interaction. Then, the target agent can adjust the searched media resources according to the supplementary information and update the media resources displayed on the user interface.

[0161] In the above embodiment, when the user is not satisfied with the searched media resources, the media resources can be updated through multiple rounds of interaction, which is conducive to further improving the accuracy of the searched media resources.

[0162] In some embodiments, the controller is further configured to: determine the media resource playback conditions and media resource playback methods based on the stored memory information through the target intelligent agent; when the current environmental information meets the media resource playback conditions, query the media resources that have not been played yet through the target intelligent agent; and trigger the continued playback of the media resources that have not been played yet according to the media resource playback method through the target intelligent agent when the user authorizes the playback.

[0163] The media resource playback condition may be the time at which the media resource is played or an event that triggers the media resource to be played. The media resource playback mode may be the volume, resolution, application program, etc. of the media resource playback. The media resource playback condition and the media resource playback mode may be searched and determined based on the memory information stored in the memory storage.

[0164] Exemplarily, when the current environmental information meets the media resource playback conditions, the target intelligent agent can query the media resources that have not been played yet, and further ask the user whether to continue playing. If the user agrees to play, the target intelligent agent can trigger the continued playback of the media resources that have not been played yet according to the media resource playback method.

[0165] For example, when watching a movie, users often set the volume between 8 and 10. They usually like to start watching movies at 8 p.m., so at 8 p.m., you can proactively check the movie history and find out if there are any unfinished movies. Then, you can proactively ask the user if they want to continue watching the movie they didn't finish yesterday. When the movie starts, you can proactively check the current volume. If it is 5, you can proactively ask or execute volume adjustment.

[0166] In this embodiment, since the target agent triggers the continued playback of the currently unplayed media resource according to the media resource playback mode stored in memory, the flexibility of media resource playback is improved, thereby improving user experience.

[0167] In some embodiments, reference Fig.13 , is a timing diagram of a method for searching for media resources based on a display device, which specifically includes the following steps:

[0168] 1. The user sends a media resource search request to the display device through voice input or text input. After receiving the user's media resource search request, the display device can determine the text input information used to describe the target task indicated by the media resource search request, and the display device can provide the text input information to the target intelligent agent installed on it.

[0169] 2. The target agent performs memory retrieval based on the input information, thereby obtaining the memory information.

[0170] 3. The target intelligent agent combines the text input information, the text input information of the target task-related memory with the corresponding basic information, the obtainable basic information list and definition, generates prompt information, and inputs the prompt information to the large model, so as to obtain the basic information list returned by the large model.

[0171] 4. The target agent searches the SOP database online based on text input information, memory information, and basic information to obtain the SOP.

[0172] 5. The target agent combines text input information, memory information, basic information, SOP and available tool description information to generate prompt information, and inputs the prompt information to the large model, so that the large model can decompose the target task into multiple subtasks. Specifically, the large model can first return the action execution instruction information of the first subtask, and the target agent calls the tool interface according to the action execution instruction information to obtain information or execute control to obtain information or execution results. Then, the target agent can prompt the large model with the obtained information or execution results again, and the large model continues to return the action execution instruction information of the next subtask. After multiple cycles, when the large model confirms that the target task is completed, the target agent can obtain at least one searched media resource.

[0173] 6. After the target agent searches for at least one media resource, it can return the media resource to the display device, so that the display device can display the searched media resource on the display.

[0174] For example, assuming the text input information is "Find me a movie starring a certain star during the 2018 Spring Festival", the process of the target agent searching for film and television resources is as follows:

[0175] 1. Use the movie and TV search tool to search for all the movies starring a certain star in 2018.

[0176] 2. Since users want movies released during the Spring Festival, they start filtering out movies and TV resources related to the Spring Festival from the search results.

[0177] 3. If found, directly return the film and television resources; if not found, try to call the film and television knowledge query tool to return relevant film and television knowledge.

[0178] 4. Call the film and television search again based on the found films and return them to the user.

[0179] Assuming the text input information is "Just find me a movie", the process of the target agent searching for film and television resources is as follows:

[0180] 1. Actively query festival / weather and other information.

[0181] 2. Today is National Day. We recommend films related to the country, the founding of the nation, the red color, and national strength.

[0182] 3. Today is a rainy day. I recommend some movies suitable for watching on rainy days, such as movies about family, warmth, love, etc.

[0183] or

[0184] 1. Query the user's movie viewing history and obtain the movie the user has watched recently;

[0185] 2. If there is a movie that you haven't finished watching, play it directly;

[0186] 3. If you have finished watching all the movies, then based on the movies you have watched, call the film and television knowledge query tool to recommend similar movies;

[0187] 4. Call the movie search to play recommended movies.

[0188] or

[0189] 1. Proactively ask users what they want to watch;

[0190] 2. Through multiple rounds of communication, we can explore users’ immediate interests and push film and television resources.

[0191] Based on the same inventive concept, the embodiment of the present application also provides a media resource search method applied to the above-mentioned display device. The implementation scheme for solving the problem provided by the method is similar to the implementation scheme recorded in the above-mentioned display device embodiment, so the specific limitations in one or more media resource search method embodiments provided below can refer to the above-mentioned limitations on the display device, and will not be repeated here.

[0192] In some embodiments, the present application further provides a media resource search method, which is applied to the above-mentioned display device. Fig.14 As shown, the media resource search method includes the following steps:

[0193] Step 1402: Receive a media resource search request from a user, and determine text input information for describing a target task indicated by the media resource search request.

[0194] Step 1404: Generate search prompt information for the target task based on the text input information.

[0195] The search prompt information includes standard operating procedures, text input information, memory information retrieved based on the text input information, and basic information for the target task. The standard operating procedures are obtained by the target agent by searching the standard operating procedure database based on the text input information, memory information, and basic information.

[0196] Step 1406: Display at least one media resource searched based on the search prompt information on the user interface.

[0197] Among them, at least one media resource is obtained by the target intelligent agent performing a subtask corresponding to the target task, and the subtask is obtained by the target intelligent agent calling the large model based on the prompt information to decompose the target task.

[0198] In some embodiments, search prompt information for a target task is generated based on text input information, including: through a target intelligent agent, memory information for the target task is obtained based on the text input information, and basic information for the target task is obtained; through a target intelligent agent, a standard operating procedure database is retrieved based on the text input information, memory information and basic information to obtain the standard operating procedure for the target task; through a target intelligent agent, text input information, memory information, basic information and standard operating procedures are spliced ​​to obtain search prompt information for the target task.

[0199] In some embodiments, basic information for a target task is obtained, including: through a target intelligent agent, determining a plurality of currently available candidate basic information and their respective definitions, and obtaining historical text input information and historical basic information related to the target task from memory information; through a target intelligent agent, inputting prompt information generated by splicing text input information, a plurality of candidate basic information and their respective definitions, historical text input information and historical basic information into a large model to obtain basic information for the target task.

[0200] In some embodiments, the target agent executes the subtask by calling the tool interface indicated by the subtask. The media resource search method also includes: through the target agent, based on text input information, retrieving at least one tool description information from a tool description information database; through the target agent, splicing the text input information, memory information, basic information and standard operating procedures to obtain search prompt information for the target task, including: through the target agent, splicing the text input information, memory information, basic information, standard operating procedures and at least one tool description information to obtain search prompt information for the target task.

[0201] In some embodiments, the media resource search method also includes: for each historical task in the historical task set, obtaining the text input information, agent execution trajectory and searched media resources corresponding to the historical task; constructing a reflective agent, and determining the execution results of the historical tasks through the reflective agent based on the text input information of the historical tasks, the agent execution trajectory and the searched media resources; filtering successful tasks whose execution results indicate successful execution from the historical task set, and calling the large model based on the filtered successful tasks to generate a first standard operating procedure; and constructing a standard operating procedure database based on the generated first standard operating procedure.

[0202] In some embodiments, a first standard operating procedure is generated by calling a large model based on the screened successful tasks, including: for each screened successful task, based on the media resource search request corresponding to each successful task, each successful task is grouped and clustered to obtain multiple cluster clusters; for each cluster cluster, from the successful tasks included in the cluster cluster, successful tasks with a number of subtasks less than a preset threshold are screened to obtain high-quality successful tasks; for each high-quality successful task, the text input information, agent thinking information and calling tool information corresponding to the high-quality successful task are spliced ​​to obtain the first prompt information, so as to prompt the large model to generate the first standard operating procedure corresponding to the high-quality successful task.

[0203] In some embodiments, the media resource search method also includes: for each non-high-quality successful task in each cluster, splicing the text input information, agent thinking information, tool calling information and tool return results corresponding to the non-high-quality successful task to obtain a second prompt information; inputting the second prompt information into the big model to prompt the big model to optimize the sub-tasks of the non-high-quality successful task, and then generate a second standard operating procedure based on the optimized sub-tasks; constructing a standard operating procedure database based on the generated first standard operating procedure, including: constructing a standard operating procedure database based on the generated first standard operating procedure and the generated second standard operating procedure.

[0204] In some embodiments, the search prompt information also includes negative prompt information retrieved from the standard operating procedure database, and the media resource search method also includes: for failed tasks, splicing the text input information, agent thinking information, tool calling information and tool return results corresponding to the failed task to obtain a third prompt information, so as to prompt the large model to determine the first subtask that has no positive effect on the failed task; for non-high-quality successful tasks, splicing the text input information, agent thinking information, tool calling information and tool return results corresponding to the failed task to obtain a fourth prompt information, so as to prompt the large model to determine the second subtask that has no positive effect on the non-high-quality successful task; generating negative prompt information based on the first subtask and the second subtask, and storing it in the standard operating procedure database.

[0205] In some embodiments, the media resource search request is a request to search for film and television resources, at least one media resource is a film and television resource, and the media resource search method also includes: controlling the display to display a preview control corresponding to each film and television resource in the user interface; in response to the user's trigger operation on the preview control, controlling the display to display a floating page in the user interface; controlling the display to play a preview clip of the film and television resource in the floating page.

[0206] In some embodiments, the media resource search method further includes: obtaining user feedback information regarding at least one media resource; when the feedback information indicates that the user is dissatisfied with at least one media resource, obtaining supplementary information input by the user; and updating the media resources displayed on the user interface according to the supplementary information.

[0207] In some embodiments, the media resource search method also includes: determining the media resource playback conditions and media resource playback methods based on the stored memory information through the target intelligent agent; when the current environmental information meets the media resource playback conditions, querying the media resources that have not been played yet through the target intelligent agent; and triggering the continued playback of the media resources that have not been played yet according to the media resource playback method after obtaining the user's authorization to play.

[0208] Based on the same inventive concept, the embodiment of the present application also provides a media resource search device for implementing the media resource search method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more media resource search device embodiments provided below can refer to the limitations on the media resource search method above, and will not be repeated here.

[0209] In some embodiments, the present application also provides a media resource search device, which is applied to the above-mentioned display device. Fig.15 As shown, in this embodiment, the media resource search device 1500 includes:

[0210] The text input information determination module 1502 is used to receive a media resource search request from a user and determine text input information used to describe a target task indicated by the media resource search request;

[0211] The search prompt information generating module 1504 is used to generate search prompt information for the target task based on the text input information; the search prompt information includes the standard operation process, the text input information, the memory information retrieved based on the text input information, and the basic information for the target task. The standard operation process is obtained by the target agent by searching the standard operation process database based on the text input information, the memory information and the basic information;

[0212] The media resource display module 1506 is used to display at least one media resource searched based on the search prompt information in the user interface; the at least one media resource is obtained by the target intelligent agent executing a subtask corresponding to the target task, and the subtask is obtained by the target intelligent agent calling the large model based on the prompt information to disassemble the target task.

[0213] In some embodiments, the search prompt information generation module 1504 is also used to: obtain memory information for the target task based on text input information through the target intelligent agent, and obtain basic information for the target task; retrieve the standard operating procedure database based on text input information, memory information and basic information through the target intelligent agent to obtain the standard operating procedure for the target task; and obtain search prompt information for the target task by splicing text input information, memory information, basic information and standard operating procedures through the target intelligent agent.

[0214] In some embodiments, the search prompt information generation module 1504 is also used to: determine, through the target intelligent agent, multiple candidate basic information and their respective definitions that are currently available, and obtain historical text input information and historical basic information related to the target task from the memory information; through the target intelligent agent, input the prompt information generated by splicing text input information, multiple candidate basic information and their respective definitions, historical text input information and historical basic information into a large model to obtain basic information for the target task.

[0215] In some embodiments, the target agent executes the subtask by calling the tool interface indicated by the subtask, and the media resource search device also includes: a tool description information search module, which is used to retrieve at least one tool description information from the tool description information database based on text input information through the target agent; a search prompt information generation module 1504, which is also used to: through the target agent, splice text input information, memory information, basic information, standard operating procedures and at least one tool description information to obtain search prompt information for the target task.

[0216] In some embodiments, the media resource search device also includes: a standard operating procedure database construction module, which is used to obtain text input information, agent execution trajectory and searched media resources corresponding to each historical task in the historical task set; construct a reflective agent, and determine the execution results of the historical tasks based on the text input information of the historical tasks, the agent execution trajectory and the searched media resources through the reflective agent; filter out successful tasks whose execution results indicate successful execution from the historical task set, and call the large model based on the filtered successful tasks to generate a first standard operating procedure; and construct a standard operating procedure database based on the generated first standard operating procedure.

[0217] In some embodiments, the standard operating procedure database construction module is also used to: for each successful task that has been screened out, based on the media resource search request corresponding to each successful task, group and cluster each successful task to obtain multiple cluster clusters; for each cluster cluster, from the successful tasks included in the cluster cluster, screen the successful tasks whose number of subtasks is less than a preset threshold to obtain high-quality successful tasks; for each high-quality successful task, splice the text input information, agent thinking information and calling tool information corresponding to the high-quality successful task to obtain the first prompt information, so as to prompt the large model to generate the first standard operating procedure corresponding to the high-quality successful task.

[0218] In some embodiments, the standard operating procedure database construction module is also used to: for each non-high-quality successful task in each cluster, splice the text input information, agent thinking information, tool calling information and tool return result corresponding to the non-high-quality successful task to obtain a second prompt information; input the second prompt information into the large model to prompt the large model to optimize the sub-tasks of the non-high-quality successful task, and then generate a second standard operating procedure based on the optimized sub-tasks; and construct a standard operating procedure database based on the generated first standard operating procedure and the generated second standard operating procedure.

[0219] In some embodiments, the search prompt information also includes negative prompt information retrieved from the standard operating procedure database, and the standard operating procedure database construction module is also used to: for failed tasks, splice the text input information, agent thinking information, tool calling information and tool return results corresponding to the failed task to obtain the third prompt information, so as to prompt the large model to determine the first subtask that has no positive effect on the failed task; for non-high-quality successful tasks, splice the text input information, agent thinking information, tool calling information and tool return results corresponding to the failed task to obtain the fourth prompt information, so as to prompt the large model to determine the second subtask that has no positive effect on the non-high-quality successful task; generate negative prompt information based on the first subtask and the second subtask, and store it in the standard operating procedure database.

[0220] In some embodiments, the media resource search request is a request to search for film and television resources, at least one media resource is a film and television resource, and the media resource search device also includes: a preview control module, used to control the display to display a preview control corresponding to each film and television resource in the user interface; in response to the user's trigger operation on the preview control, control the display to display a floating page in the user interface; control the display to play a preview clip of the film and television resource in the floating page.

[0221] In some embodiments, the media resource search device also includes: a media resource update module, used to obtain user feedback information on at least one media resource; when the feedback information indicates that the user is dissatisfied with at least one media resource, obtain supplementary information input by the user; and update the media resources displayed on the user interface according to the supplementary information.

[0222] In some embodiments, the media resource search device also includes: a playback trigger module, which is used to determine the media resource playback conditions and the media resource playback method through the target intelligent agent based on the stored memory information; when the current environmental information meets the media resource playback conditions, query the media resources that have not been played yet through the target intelligent agent; and when the user authorizes the playback, trigger the continued playback of the media resources that have not been played yet through the target intelligent agent according to the media resource playback method.

[0223] Each module in the above media resource search device can be implemented in whole or in part by software, hardware or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute operations corresponding to each module.

[0224] In some embodiments, a display device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and the above method steps are implemented when the processor executes the computer program.

[0225] In some embodiments, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the above method steps are implemented.

[0226] In some embodiments, a computer program product is provided, comprising a computer program, which implements the above method steps when executed by a processor.

[0227] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0228] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0229] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A display device, characterized in that: The device comprises: a display configured to display content from a broadcast system or network and / or a user interface, and to display a user interface; The controller is configured as: Receive a media resource search request from a user, and determine text input information for describing a target task indicated by the media resource search request; Based on the text input information, generate search prompt information for the target task; the search prompt information includes a standard operating procedure, the text input information, memory information retrieved based on the text input information, and basic information for the target task, and the standard operating procedure is obtained by the target agent by searching a standard operating procedure database based on the text input information, the memory information, and the basic information; Control the display to display in the user interface at least one media resource searched based on the search prompt information; the at least one media resource is obtained by the target agent executing a subtask corresponding to the target task, and the subtask is obtained by the target agent calling a large model based on the search prompt information to decompose the target task.

2. The device according to claim 1, characterized in that When the controller executes the step of generating search prompt information for the target task based on the text input information, the controller is further configured to: Acquiring, by the target agent, memory information for the target task based on the text input information, and acquiring basic information for the target task; Through the target agent, based on the text input information, the memory information and the basic information, a standard operating procedure database is retrieved to obtain a standard operating procedure for the target task; Through the target intelligent agent, the text input information, the memory information, the basic information and the standard operating procedure are spliced ​​together to obtain search prompt information for the target task.

3. The device according to claim 2, characterized in that When the controller executes the step of acquiring basic information for the target task, the controller is further configured to: Determine, through the target agent, a plurality of candidate basic information currently available and their respective definitions, and obtain historical text input information and historical basic information related to the target task from the memory information; Through the target intelligent agent, the prompt information generated by splicing the text input information, the multiple candidate basic information and their respective definitions, the historical text input information and the historical basic information is input into the large model to obtain the basic information for the target task.

4. The device according to claim 2, characterized in that The target agent executes the subtask by calling the tool interface indicated by the subtask, and the controller is further configured to: Retrieving, by the target agent, at least one tool description information from a tool description information database based on the text input information; When the controller executes the step of obtaining search prompt information for the target task by splicing the text input information, the memory information, the basic information and the standard operation process through the target agent, the controller is further configured as follows: Through the target intelligent agent, the text input information, the memory information, the basic information, the standard operating procedure and the at least one tool description information are spliced ​​together to obtain search prompt information for the target task.

5. The device according to claim 1, characterized in that The controller is also configured to: For each historical task in the historical task set, obtain text input information, agent execution trajectory, and searched media resources corresponding to the historical task; Constructing a reflective agent, and determining the execution result of the historical task through the reflective agent based on the text input information of the historical task, the agent execution trajectory and the searched media resources; Filtering successful tasks whose execution results indicate successful execution from the historical task set, and calling the large model to generate a first standard operating procedure based on the filtered successful tasks; The standard operating procedure database is constructed based on the generated first standard operating procedure.

6. The device according to claim 5, characterized in that When the controller executes the step of generating a first standard operating procedure by calling the large model based on the screened successful tasks, the controller is further configured to: For each successful task that has been screened out, based on a media resource search request corresponding to each successful task, each successful task is grouped and clustered to obtain a plurality of clusters; For each cluster, from the successful tasks included in the cluster, select successful tasks whose number of subtasks is less than a preset threshold to obtain high-quality successful tasks; For each high-quality successful task, the text input information, agent thinking information and calling tool information corresponding to the high-quality successful task are spliced ​​to obtain the first prompt information to prompt the large model to generate the first standard operating procedure corresponding to the high-quality successful task.

7. The device according to claim 6, characterized in that The controller is also configured to: For each non-high-quality successful task in each cluster, the text input information, agent thinking information, tool calling information and tool return result corresponding to the non-high-quality successful task are spliced ​​to obtain the second prompt information; Inputting the second prompt information into the large model to prompt the large model to optimize the subtasks of the non-high-quality successful task and generate a second standard operating procedure based on the optimized subtasks; When the controller constructs the standard operating procedure database based on the generated first standard operating procedure, the controller is further configured to: The standard operating procedure database is constructed based on the generated first standard operating procedure and the generated second standard operating procedure.

8. The device according to claim 6, characterized in that The search prompt information also includes negative prompt information retrieved from the standard operating procedure database, and the controller is further configured to: For the failed task, the text input information, agent thinking information, tool calling information and tool return result corresponding to the failed task are spliced ​​to obtain the third prompt information, so as to prompt the large model to determine the first subtask that has no positive effect on the failed task; For non-high-quality successful tasks, the text input information, agent thinking information, tool calling information and tool return result corresponding to the failed task are spliced ​​to obtain fourth prompt information to prompt the large model to determine the second subtask that has no positive effect on the non-high-quality successful task; Negative prompt information is generated based on the first subtask and the second subtask, and stored in the standard operating procedure database.

9. The device according to any one of claims 1 to 8, characterized in that The media resource search request is a request for searching for film and television resources, the at least one media resource is a film and television resource, and the controller is further configured to: Control the display to display a preview control corresponding to each film and television resource in the user interface; In response to a trigger operation of the user on the preview control, controlling the display to display a floating page on the user interface; The display is controlled to play a preview clip of the film and television resource in the floating page.

10. The device according to any one of claims 1 to 8, characterized in that The controller is also configured to: Acquiring user feedback information regarding the at least one media resource; When the feedback information indicates that the user is dissatisfied with the at least one media resource, obtaining supplementary information input by the user; The media resources displayed on the user interface are updated according to the supplementary information.

11. The device according to any one of claims 1 to 8, characterized in that The controller is also configured to: Determining media resource playback conditions and media resource playback methods based on stored memory information through the target agent; When the current environment information meets the media resource playback conditions, query the media resources that are not currently played through the target agent; In the case of obtaining the user's authorization to play, the target agent triggers the continued playing of the media resource that is not yet played according to the media resource playing method.

12. A media resource search method, characterized in that: The method comprises: Receive a media resource search request from a user, and determine text input information for describing a target task indicated by the media resource search request; Based on the text input information, generate search prompt information for the target task; the search prompt information includes a standard operating procedure, the text input information, memory information retrieved based on the text input information, and basic information for the target task, and the standard operating procedure is obtained by the target agent by searching a standard operating procedure database based on the text input information, the memory information, and the basic information; On the user interface, at least one media resource searched based on the search prompt information is displayed; the at least one media resource is obtained by the target intelligent agent executing a subtask corresponding to the target task, and the subtask is obtained by the target intelligent agent calling a large model based on the prompt information to disassemble the target task.

Images